Hi all!

I wanted to share a little bit about my experience during the 2026 ME72 Competition: the largest Mechanical Engineering competition within Caltech, where we build robots and compete.

As the capstone project for Mechanical Engineering students, ME72 has definitely been one of the most rewarding experiences I’ve had at Caltech. It was challenging, stressful, and very time-consuming, but also one of the most fun classes I’ve taken.

(You can also look at Diya’s blog post for another perspective about this year’s competition!)

ME72 is a two-term class that serves as the Mechanical Engineering capstone. Students can choose to complete a senior thesis (ME90) instead, but most people choose ME72 because it offers the opportunity to design, build, and compete with a robot from scratch. It’s a huge commitment, and everyone warns you that the class will take over your life for two terms. Unfortunately, it’s true, but totally worth it.

Every year the competition changes, and this year’s game was called Apex Cleanup: Summit, Mint, Bank. One of our teammates helped design the art for the poster.

The game had multiple components. Each round started with a 30-second autonomous period, where one robot could press a button for bonus points while the second robot could begin collecting pellets. After autonomous, drivers took over and the main objective became collecting pellets, climbing a four-foot pyramid, and depositing the pellets at the top for points. There were also smaller two-foot depots assigned to each team where pellets could be deposited if a robot couldn’t make the climb. However, these were worth only one-third of the points. There’s more stuff to it, so feel free to check more specifics on the ME72 website.

At first glance, it might sound simple, but it was so much harder than it looked. Designing a robot for this competition needed so many different components like climbing clearance, climbing ability, intake design, speed, weight, magnet addition, and maneuverability, all while staying within a total budget of $1200 for up to two robots.

Fall term was mostly dedicated to the design process. Within just a few weeks we had to complete our Preliminary Design Review (PDR) and then our Critical Design Review (CDR), so there wasn’t much time to waste. My team met two or three times every week to discuss designs, materials, budgets, and scheduling.

Looking back, I think one of the hardest parts was actually deciding what to buy. With such a limited budget, every purchase mattered, and making a mistake could be expensive.

Fortunately, our team spent a lot of time doing calculations and thinking through our decisions before ordering parts. That ended up being one of our biggest strengths, since we never had to completely redesign the robot later on. Of course, after every test we had to iterate. I remember changing our wheels was one thing we had to do that caused a lot of stress, since they were initially so expensive.

By the end of fall term, the goal was to have one fully functional drivetrain, including the chassis, motors, and much of the electronics. It was only after repeated testing and our demonstration that we realized our wheels weren’t going to survive climbing the pyramid. The torque was so large that the plastic hubs detached from the tires, so we had to find new wheels.

Our team had seven people, which was relatively large compared to previous years. During fall term, I mainly worked on the drivetrain, but over winter term I shifted toward electronics and autonomous programming because I wanted to gain more experience with that side of the robot, and I was going to be the one implementing the autonomous portion since I wanted to use line-follow logic. At the same time, I still helped with machining and drivetrain work whenever needed.

So this is what our robots looked like, and as you can probably tell from the pictures, there was a lot of 3D printing involved. That decision was mostly due to the budget. While 3D-printed parts are not the best option for components that experience repeated impacts and heavy loads, they save us a huge amount of machining time and money. 

However, many of our failures throughout the year came because of those 3D printed parts, so we spent a lot of time repairing components and printing replacements. I remember we had to print so many wheel hubs and so many magnet holders because for some reason we decided to print those even though they experienced some of the heavier forces (it was for money reasons probably). We just knew that during competition, we’d have to take extra parts. Thankfully, the design held together long enough to get us through the tournament. Although we did have to change out some of our wheel hubs right before our semifinal round.

Now onto competition day.

This was easily one of the most exciting days of the year. At the time it was incredibly stressful because so many people from across campus come to watch, and the last thing you want is for your robot to stop working in front of everyone, although it does happen all the time. But looking back, it was also one of the most rewarding experiences I’ve had at Caltech.

Our team was called the PharaoBots, and our two robots were named Ramses and Cleopatra.

Competition day was actually the first time we could officially test our autonomous code on the real field. Throughout the term we had implemented line-following logic, but we never had access to the actual competition field beforehand. A couple of us showed up early that morning to tune sensors and test our code before the matches started.

The autonomous period only lasted thirty seconds, but it was one of the easiest ways to score points. One robot raced to press a button while the other immediately started collecting pellets. Thankfully, this ended up being one of our strongest parts of the competition as we were constantly first at pressing the button.

Then came the climbing.

Daniel and I had worked on much of the electronics and driver controls, so we knew all the driving modes. We had programmed different driving modes and controller mappings, and we didn’t find time for everyone on the team to learn them before competition day, they also didn’t feel as comfortable. As a result, Daniel and I ended up driving throughout the entire tournament.

We were also allowed one teammate to stand on a black box for a better view of the field. That ended up being Anya, who communicated with us through headphones and helped direct us since we couldn’t really see the top.

• That is Anya sitting on the black box. During the round, she would stand on it and have a clearer view of what was happening at the top of the pyramid.

As the rounds went on, we also learned a lot about game strategy. Every match taught us something new about where to drive, which pellets to prioritize, and how to coordinate the two robots.

I always drove Ramses, which focused on the autonomous button, while Daniel drove Cleopatra, our stronger climber. We aimed for each robot to climb at least once per round, and ideally Cleopatra would climb twice. However, by the end of the day, Ramses was the only one climbing, so we had to change our strategy so that Clepatra would focus more on scoring on the smaller depots. Fortunately, we were one of the few teams still climbing successfully by the end of the competition.

Ultimately, we won.

It was honestly one of the best feelings I’ve had during my time at Caltech. Going into the competition, our biggest hope was simply that the robots would work. We never expected to win. But as the rounds went on and we kept improving our driving and strategy, we started scoring more and more points until suddenly winning felt possible.

Looking back, I think ME72 taught me much more than just how to build a robot. It taught me how to work on a large team, manage a budget, troubleshoot problems under pressure, and adapt when things don’t go according to plan. It was one of the most challenging classes I’ve taken, but also one of the most rewarding.

We don’t get to keep the giant trophy, but we do get tiny trophies with our names on them, which I think are cute.

More pics: